Primitive neuroectodermal tumors (PNET), including medulloblastoma (MB), are the most common malignant brain tumors of children. To approach the unsolved problem of effective therapy, the development and progression of PNET/MB must first be understood. We hypothesize that the shedding of structurally novel tumor cell surface gangliosides influences the biology of PNET/MB. To begin to test this hypothesis, we propose to fully characterize the structures and metabolism of PNET/MB gangliosides. To accomplish these goals, we shall separate and purify PNET/MB gangliosides into individual homogeneous molecular species using newly developed HPLC purification methodology. Using FAB and MALDI-MS, we shall elucidate their complete carbohydrate and ceramide structures. Furthermore, we shall chemically detect tumor ganglioside shedding into the cerebrospinal fluid, a site of frequent tumor metastasis in this brain tumor. Finally, we shall use pharmacologic agents (e.g., retinoic acid) to probe and possibly alter ganglioside metabolism (both synthesis and shedding) of PNET/MB cells. These studies are directed to the ultimate goals of understanding the role of PNET/MB gangliosides in tumor development, and of modulating their expression. A library of novel gangliosides provided by this proposed work will allow fulfillment of longer-term aims of determination of the biological activity of specific PNET/MB gangliosides, definition of their role in enhancing brain tumor formation, and development of new monoclonal antibodies for diagnosis and treatment of PNET/MB. The overall project should greatly increase our understanding of PNET/MB, and may facilitate development of novel therapeutic approaches such as biochemical eingineering of the tumor cell to alter ganglioside synthesis and shedding.